Top AI Agents Built to Catch Malicious Code Can Be Tricked Into Running It
Ask an AI coding agent to scan open-source code for security holes, and it might run the attacker's code on your own machine instead
Open-source software enables code review and reuse, but known vulnerabilities and unmaintained dependencies can create cybersecurity risks.
Search across headline titles and summaries.
Background for this topic.
Open source is software whose source code is available under a license that permits use, inspection, modification, and redistribution. It may be developed by a community, an organization, or a small group of maintainers; “open” does not guarantee that the code is actively reviewed, supported, or secure.
For security teams, the main concerns are vulnerabilities in dependencies and the software supply chain: a maintainer account, release process, or package can be compromised, while an unmaintained component may retain known flaws. Public code can enable review and faster fixes, but visibility alone is not a control. Maintain an inventory or SBOM of open-source components, pin and verify versions or signatures where possible, monitor vulnerability advisories, and apply updates through a controlled process.
Weekly headline count for the current query.
Ask an AI coding agent to scan open-source code for security holes, and it might run the attacker's code on your own machine instead
Cybersecurity researchers have described what they say is a new class of attack that can trick artificial intelligence (AI) coding agents into running arbitrary code on developer machines